Mérouane Debbah, Mathematical and Algorithmic Sciences Lab, Huawei, France


Random Matrix Methods for Wireless Communications


The following talk aims to provide a comprehensive understanding on how random matrix theory can model the complexity of the interaction between wireless devices. One of the useful features of the large dimensional random matrix theory approach, is its ability to predict, under certain conditions, the behavior of the empirical eigenvalue distribution of products and sums of matrices. The results are striking in terms of accuracy compared to simulations with reasonable matrix sizes, and the theory has been shown to be an efficient tool to predict the behavior of wireless systems with only few meaningful parameters. In this talk, we will introduce the recent results in the field, in particular the "Deterministic Equivalent" approach and its application to the design of Massive MIMO systems.